Assembly bias in quadratic bias parameters of dark matter halos from forward modeling

We use the forward modeling approach to galaxy clustering combined with the likelihood from the effective-field theory of large-scale structure to measure assembly bias, i.e. the dependence of halo bias on properties beyond the total mass, in the linear (b1) and second order bias parameters (b2 and bK2) of dark matter halos in N-body simulations. This is the first time that assembly bias in the tidal bias parameter bK2 is measured. We focus on three standard halo properties: the concentration c, spin λ, and sphericity s, for which we find an assembly bias signal in bK2 that is opposite to that in b1. Specifically, at fixed mass, halos that get more (less) positively biased in b1, get less (more) negatively biased in bK2. We also investigate the impact of assembly bias on the b2(b1) and bK2(b1) relations, and find that while the b2(b1) relation stays roughly unchanged, assembly bias strongly impacts the bK2(b1) relation. This impact likely extends also to the corresponding relation for galaxies, which motivates future studies to design better priors on bK2(b1) for use in cosmological constraints from galaxy clustering data.